BlogLLM Benchmark Scores as Published, Side by Side
DeepSWE: Results as Published
Every result the catalog holds for DeepSWE: 58 models from 14 developers, read from 34 pages. Each number is as printed by the company named beside it; nothing is ranked or rescored. Models are listed by developer, newest first.
Area Agentic CodingMaintainer DatacurveIntroduced 2026Data version 2026103001, as of 2026-10-30
DeepSWE tests whether coding agents can make original, long-horizon software engineering changes in real open-source repositories. It has 113 tasks across 91 repositories in TypeScript, Go, Python, JavaScript and Rust, all written from scratch so that solutions are not copied from existing commits. Hand-written verifiers check behavior through public interfaces, and each verifier runs three times to catch flakiness.
DeepSWE v1.1
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPT-6 LunaOpenAI, released 2026-09-22 | 66.6% | OpenAI, Introducing GPT-6 Sol and GPT-6 Lunachecked 2026-10-19 | Effort: max |
| GPT-6 SolOpenAI, released 2026-09-22 | 68.8% | OpenAI, Introducing GPT-6 Sol and GPT-6 Lunachecked 2026-10-19 | Effort: max |
| GPT-6 AstraOpenAI, released 2026-09-03 | 74.1% | OpenAI, GPT-6 Astrachecked 2026-10-14; printed on 3 more pages | - |
| GPT-5.6 LunaOpenAI, released 2026-07-09 | 67.2% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15; printed on 1 more page | - |
| GPT-5.6 SolOpenAI, released 2026-07-09 | 72.7% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15; printed on 13 more pages | - |
| GPT-5.6 TerraOpenAI, released 2026-07-09 | 69.6% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15; printed on 3 more pages | - |
| GPT-5.5OpenAI, released 2026-04-23 | 67% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15; printed on 3 more pages | - |
| Claude Sonnet 5.5Anthropic, released 2026-09-28 | 71% | Anthropic, Claude Sonnet 5.5 System Cardchecked 2026-10-19 | - |
| Claude Opus 5.5Anthropic, released 2026-09-22 | 74.2% | Anthropic, Claude Opus 5.5 System Cardchecked 2026-10-25; printed on 3 more pages | - |
| Claude Fable 5.1Anthropic, released 2026-09-01 | 67.4% | Anthropic, System Card: Claude Fable 5.1 & Claude Mythos 5.1checked 2026-10-25; printed on 5 more pages | - |
| Claude Opus 5Anthropic, released 2026-07-24 | 68.8% | Anthropic, Claude Opus 5checked 2026-10-15; printed on 6 more pages | - |
| Claude Sonnet 5Anthropic, released 2026-06-30 | 54% | Google, Gemini 3.6 Flash model cardchecked 2026-10-15; printed on 3 more pages | - |
| Claude Fable 5Anthropic, released 2026-06-09 | 69.7% | Anthropic, Claude Opus 5checked 2026-10-15; printed on 11 more pages | - |
| Claude Opus 4.8Anthropic, released 2026-05-28 | 59% | Anthropic, Claude Opus 5checked 2026-10-15; printed on 9 more pages | - |
| Gemini 4 ArgonGoogle, released 2026-09-30 | 77.9% | Google, Gemini 4 Argonchecked 2026-10-29 | - |
| Gemini 3.8 FlashGoogle, released 2026-09-02 | 73.7% | Google, Gemini 3.8 Flash model cardchecked 2026-10-15; printed on 1 more page | - |
| Gemini 3.7 FlashGoogle, released 2026-08-13 | 65.3% | Google, Gemini 3.7 Flash model cardchecked 2026-10-15; printed on 2 more pages | - |
| Gemini 3.6 FlashGoogle, released 2026-07-21 | 49% | Google, Gemini 3.6 Flash model cardchecked 2026-10-15; printed on 1 more page | - |
| Gemini 3.5 FlashGoogle, released 2026-05-19 | 37% | Google, Gemini 3.6 Flash model cardchecked 2026-10-15 | - |
| Gemini 3.1 Pro PreviewGoogle, released 2026-02-19 | 12% | Google, Gemini 3.6 Flash model cardchecked 2026-10-15; printed on 2 more pages | - |
| Grok 4.7xAI, released 2026-09-21 | 71% | xAI, Grok 4.7checked 2026-10-25; printed on 1 more page | Effort: high |
| Grok 4.6xAI, released 2026-08-12 | 65.9% | xAI, Introducing Grok 4.6checked 2026-10-15; printed on 2 more pages | Effort: high |
| Grok 4.5xAI, released 2026-07-16 | 53% | xAI, Grok 4.5checked 2026-10-15; printed on 2 more pages | Harness: mini-swe-agent (Datacurve run) |
| Muse Spark OpenMeta, released 2026-10-15 | 60% | Meta, Meta Muse Spark Open 816B-A42B model cardchecked 2026-10-19 | Effort: xhigh; Harness: preferred harness of each model |
| Muse WonderMeta, released 2026-10-12 | 78.8% | Meta, Meta Muse Wonder model cardchecked 2026-10-19 | Effort: max |
| Muse Wonder FullMeta, released 2026-10-12 | 79% | Meta, Meta Muse Wonder model cardchecked 2026-10-19 | Effort: max |
| Muse Spark 1.2Meta, released 2026-08-05 | 54.9% | Google, Gemini 3.7 Flash model cardchecked 2026-10-15 | - |
| DeepSeek-V4.1-Flash-1017DeepSeek, released 2026-10-17 | 77% | DeepSeek, DeepSeek-V4.1-Flash-1017 model cardchecked 2026-10-18; printed on 2 more pages | Harness: Mini-SWE-V2 |
| DeepSeek-V4.1-FlashDeepSeek, released 2026-09-10 | 69.8% | DeepSeek, DeepSeek-V4.1-Flash model cardchecked 2026-10-15; printed on 2 more pages | Effort: max; Harness: Claude Code; Attempt budget: N=8 samples per task; Context: 1M |
| 74.2% | DeepSeek, DeepSeek-V4.1-Flash model cardchecked 2026-10-15 | Effort: max; Harness: mini-SWE; Attempt budget: N=8 samples per task (scaffold table); Context: 1M | |
| 72.6% | DeepSeek, DeepSeek-V4.1-Flash model cardchecked 2026-10-15 | Effort: max; Harness: DeepSeek Harness (DSH) Minimal; Attempt budget: N=8 samples per task; Context: 1M | |
| 65.6% | DeepSeek, DeepSeek-V4.1-Flash model cardchecked 2026-10-15 | Effort: max; Harness: Codex; Attempt budget: N=8 samples per task; Context: 1M | |
| 66.2% | DeepSeek, DeepSeek-V4.1-Flash model cardchecked 2026-10-15 | Effort: max; Harness: Pi; Attempt budget: N=8 samples per task; Context: 1M | |
| 70.5% | DeepSeek, DeepSeek-V4.1-Flash model cardchecked 2026-10-15 | Effort: max; Harness: DeepSeek Harness (DSH) Standard; Attempt budget: N=8 samples per task; Context: 1M | |
| 67.6% | DeepSeek, DeepSeek-V4.1-Flash model cardchecked 2026-10-15 | Effort: max; Harness: DeepSeek Harness (DSH) PTC; Attempt budget: N=8 samples per task; Context: 1M | |
| 65.5% | DeepSeek, DeepSeek-V4.1-Flash model cardchecked 2026-10-15 | Effort: max; Harness: OpenCode; Attempt budget: N=8 samples per task; Context: 1M | |
| DeepSeek-V4-Flash (Preview)DeepSeek, released 2026-04-24 | 7.3% | DeepSeek, DeepSeek-V4-Flash-0731 model cardchecked 2026-10-15; printed on 1 more page | - |
| DeepSeek-V4-Pro (Preview)DeepSeek, released 2026-04-24 | 12.8% | DeepSeek, DeepSeek-V4-Flash-0731 model cardchecked 2026-10-15; printed on 2 more pages | - |
| DeepSeek-V4-Flash-0731DeepSeek | 54.4% | DeepSeek, DeepSeek-V4-Flash-0731 model cardchecked 2026-10-15; printed on 4 more pages | - |
| DeepSeek-V4-Pro-0813DeepSeek | 62.7% | DeepSeek, DeepSeek-V4-Pro-0813 model cardchecked 2026-10-15; printed on 3 more pages | - |
| DeepSeek-V4-Vision-ExpDeepSeek | 59.3% | Z.ai, Z.ai GLM-5.3-Flash model cardchecked 2026-10-15 | Harness: mini-swe-agent |
| Qwen3.8-MaxAlibaba, released 2026-08-03 | 56.6% | Alibaba, Qwen3.8-Max comparisonchecked 2026-10-14; printed on 2 more pages | - |
| Qwen3.7-PlusAlibaba, released 2026-06-01 | 14.2% | Alibaba, Qwen3.8-27B model cardchecked 2026-10-15; printed on 1 more page | Harness: Claude Code |
| Qwen3.6-27BAlibaba, released 2026-04-22 | 13.3% | Alibaba, Qwen3.8-27B model cardchecked 2026-10-15 | Harness: Claude Code |
| Qwen3.5-122B-A10BAlibaba, released 2026-02-24 | 0.4% | NVIDIA, NVIDIA Nemotron 3.5 Super VL 120B-A12B model cardchecked 2026-10-19 | - |
| Qwen3.7-MaxAlibaba | 21.6% | Alibaba, Qwen3.8-Max comparisonchecked 2026-10-14; printed on 1 more page | - |
| Qwen3.8-2.4T-A95BAlibaba | 57% | Meta, Meta Muse Spark Open 816B-A42B model cardchecked 2026-10-19 | Effort: xhigh; Harness: preferred harness of each model |
| Qwen3.8-27BAlibaba | 42.2% | Alibaba, Qwen3.8-27B model cardchecked 2026-10-15; printed on 1 more page | Harness: Claude Code |
| Qwen3.8-Flash-NextAlibaba | 58.7% | Alibaba, Qwen3.8-Flash-Next model cardchecked 2026-10-15; printed on 1 more page | Harness: best of Claude Code and mini-SWE-agent; Context: 256000 |
| Kimi K3.1 ProMoonshot AI, released 2026-10-23 | 75.5% | Moonshot AI, Introducing Kimi K3.1 Prochecked 2026-10-24 | Effort: max; Harness: Claude Code; Dataset revision: v1.1 |
| Kimi K3Moonshot AI, released 2026-07-16 | 67.5% | Moonshot AI, Kimi K3 model cardchecked 2026-10-15; printed on 7 more pages | Effort: max; Harness: Kimi Code |
| 67.3% | Moonshot AI, Kimi K3 model cardchecked 2026-10-15 | Effort: max; Harness: mini-SWE-agent | |
| GLM-5.3-FlashZ.ai, released 2026-08-26 | 63.4% | Z.ai, Z.ai GLM-5.3-Flash model cardchecked 2026-10-15; printed on 1 more page | Effort: max; Harness: mini-swe-agent |
| GLM-5.3Z.ai, released 2026-08-18 | 66.9% | Z.ai, GLM-5.3 model cardchecked 2026-10-15; printed on 6 more pages | Effort: max; Harness: mini-swe-agent; Context: 400000 |
| GLM-5.2Z.ai, released 2026-06-16 | 46.2% | Z.ai, GLM-5.2 model cardchecked 2026-10-15; printed on 6 more pages | Harness: mini-swe-agent; Context: 400000 |
| GLM-5.1Z.ai, released 2026-04-07 | 18% | Z.ai, GLM-5.2 model cardchecked 2026-10-15 | Harness: mini-swe-agent; Context: 400000 |
| Mistral Large 4 (preview)Mistral AI, released 2026-10-06 | 61.7% | Mistral AI, Mistral Large 4 (preview)checked 2026-10-15; printed on 1 more page | Harness: Mistral Vibe |
| MiniMax-M3.1-FlashMiniMax, released 2026-10-19 | 73.7% | MiniMax, MiniMax-M3.1-Flash model cardchecked 2026-10-25; printed on 1 more page | Harness: mini-SWE-agent 2.4.5; Dataset revision: v1.1 |
| MiniMax M3MiniMax, released 2026-06-01 | 20% | Z.ai, GLM-5.2 model cardchecked 2026-10-15 | Harness: mini-swe-agent; Context: 400000 |
| Nemotron 3.5 Super VL 120B-A12BNVIDIA, released 2026-10-15 | 24.65% | NVIDIA, NVIDIA Nemotron 3.5 Super VL 120B-A12B model cardchecked 2026-10-19 | - |
| Nemotron 3 UltraNVIDIA, released 2026-06-04 | 3% | NVIDIA, NVIDIA Nemotron 3.5 Super VL 120B-A12B model cardchecked 2026-10-19 | - |
| MiMo-V2.6-FlashXiaomi, released 2026-09-21 | 67.9% | Xiaomi, Xiaomi MiMo-V2.6-Pro-RL model cardchecked 2026-10-19 | - |
| MiMo-V2.6-ProXiaomi, released 2026-09-21 | 71.9% | Xiaomi, Xiaomi MiMo-V2.6-Pro-RL model cardchecked 2026-10-19 | - |
| MiMo-V2.5-ProXiaomi, released 2026-04-27 | 19% | Xiaomi, Xiaomi MiMo-V2.6-Pro-RL model cardchecked 2026-10-19 | - |
| Inkling-SmallThinking Machines | 18% | NVIDIA, NVIDIA Nemotron 3.5 Super VL 120B-A12B model cardchecked 2026-10-19 | - |
DeepSWE v1.0
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPT-5.5OpenAI, released 2026-04-23 | 64.31% | xAI, Grok 4.5checked 2026-10-15 | Effort: xhigh; Harness: each provider's own harness (Artificial Analysis run) |
| Claude Fable 5Anthropic, released 2026-06-09 | 66.1% | xAI, Grok 4.5checked 2026-10-15 | Effort: max; Harness: each provider's own harness (Artificial Analysis run) |
| Claude Opus 4.8Anthropic, released 2026-05-28 | 55.75% | xAI, Grok 4.5checked 2026-10-15 | Effort: max; Harness: each provider's own harness (Artificial Analysis run) |
| Claude Opus 4.7Anthropic, released 2026-04-16 | 40.12% | xAI, Grok 4.5checked 2026-10-15 | Effort: max; Harness: each provider's own harness (Artificial Analysis run) |
| Grok 4.5xAI, released 2026-07-16 | 62% | xAI, Grok 4.5checked 2026-10-15 | Harness: each provider's own harness (Artificial Analysis run) |
Every number here is read from the same catalog file as the benchmark table, data version 2026103001, and is the number the table shows for that model and test. Where one page printed a model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.